Molecular Landscape of MDS and its clinical …...Molecular Landscape of MDS and its clinical applications Papaemmanuil E, PhD Assistant Attending Computational Oncology MSKCC Disclosure

Molecular Landscape of MDS and its clinical applications

Papaemmanuil E, PhD

Assistant Attending

Computational Oncology

MSKCC

Disclosure

I have no relevant financial relationships to disclose.

Mye

lod

ysp

last

ic s

ynd

rom

es

The post cancer gene discovery promise

• Biology;

• Molecular oncology;

• Patient tailored medicine;

• Development of rational clinical and therapeutic protocols;

• Well – documented gene mutations, frequencies and prevelance amongst the key MDS subtypes;

• Increasingly understood patterns of co-mutation;

• Recent insights into the biological mechanisms of spliceosome deregulation and downstream

effectors;

Mye

lod

ysp

last

ic s

ynd

rom

es

Papaemmanuil et al, Blood 2013 Hafferlach et al, Leukemia 2014

Mye

lod

ysp

last

ic s

ynd

rom

es

Incorporation into diagnostic and clinical practices is not yet clear;

The MDS genome - the challenge

1. >200 recurrently mutated genes;2. Mostly infrequently mutated genes <5%;3. Most patients have more than one mutation;4. Clonal heterogeneity;5. Systematic – pairwise interactions; 6. Increasing complexity in variables to be considered in clinical practice;

Our goal…

1. Insights into the molecular mechanisms of disease biology;

1. Diagnostic biomarkers – genotype, phenotype correlations;

2. Prognostic biomarkers – clinical outcome relationships;

3. Predictive biomarkers – response to therapeutic intervention;

4. Molecular tools to monitor disease progression and response;

5. Therapies tailored to the individual genetic and clonal profiles;

Mye

lod

ysp

last

ic s

ynd

rom

es

Mye

lod

ysp

last

ic s

ynd

rom

es

Splic

ing

fact

or

mu

tati

on

s ar

e d

isea

se d

efin

ing

Splicing factors are disease defining, and much of the focus of biomarker research in MDS

Mye

lod

ysp

last

ic s

ynd

rom

es

SF3

B1

:MD

S Mutations in SF3B1 define a distinct molecular and clinical subgroup in MDS

• Lower blast counts; • Higher presence of ringed sideroblasts; • Lower incidence of multi lineage dysplasia (46.3% vs 82.6%, P < .001); • Lower proportion of dysplastic myeloid cells and megakaryocytes;• Higher absolute neutrophil and platelet counts (P < .001); • Better overall survival (P= 0.003) and lower cumulative incidence of disease progression

Patterns of co-mutation

• SF3B1 mutations define a distinct molecular subgroup that is mutually exclusive tomutations in TP53, and / or cytogenetic abnormalities or complex karyotype.

• Co-mutation with DNMT3A (20%) : - No effect on overall survival or event free survival - Increased involvement of multi lineage dysplasia

• Co-mutation with ASXL1 (7%), RUNX1 (6%) , EZH2 (5%) : - Increased transfusion dependency - Worse prognosis ( P= 0.004), disease progression (P=0.002)

Larger numbers are warranted to study the effects of co-mutation in MDS prognosisMalcovati Blood 2015

SF3

B1

:MD

S 1. Re-evaluation of bone marrow morphology in patients with SF3B1 mutations but

not diagnosed as an MDS with ringed sideroblasts, identifies that SF3B1 predicts for the presence of RS.

2. We show the % of RS is correlated with % VAF of SF3B1 mutations

Malcovati Blood 2012;

Sf3B1 mutation 97% positive predictive value for ringed sideroblastsSF3

B1

:MD

S

SF3

B1

:MD

S 1. Re-evaluation of bone marrow morphology in patients with SF3B1 mutations but

not diagnosed as an MDS with ringed sideroblasts, identifies that SF3B1 predicts for the presence of RS.

2. We show the % of RS is correlated with % VAF of SF3B1 mutations

Sf3B1 mutation 97% positive predictive value for ringed sideroblasts

SF3

B1

:MD

S

Malcovati Blood 2012;

SF3

B1

:MD

S Mutations in SF3B1 define a distinct molecular and clinical subgroup in MDS

• Lower blast counts • Higher presence of ringed sideroblasts • Lower incidence of multi lineage dysplasia (46.3% vs 82.6%, P < .001), • Lower proportion of dysplastic myeloid cells and megakaryocytes• Higher absolute neutrophil and platelet counts (P < .001),

• Better overall survival (P= 0.003) and lower cumulative incidence of disease progression

Patterns of co-mutation

• SF3B1 mutations define a distinct molecular subgroup that is mutually exclusive tomutations in TP53, and / or cytogenetic abnormalities or complex karyotype.

• Co-mutation with DNMT3A (20%) : - No effect on overall survival or event free survival; - Increased involvement of multi lineage dysplasia

• Co-mutation with ASXL1 (7%), RUNX1 (6%) , EZH2 (5%) : - Increased transfusion dependency - Worse prognosis ( P= 0.004), disease progression (P=0.002)

Larger numbers are warranted to study the effects of co-mutation in MDS prognosis

SF3

B1

:MD

S

Malcovati Blood 2015

SF3

B1

:MD

S

SF3

B1

:MD

S Mutations in SF3B1 define a distinct molecular and clinical subgroup in MDS

• Lower blast counts • Higher presence of ringed sideroblasts • Lower incidence of multi lineage dysplasia (46.3% vs 82.6%, P < .001), • Lower proportion of dysplastic myeloid cells and megakaryocytes• Higher absolute neutrophil and platelet counts (P < .001),

• Better overall survival (P= 0.003) and lower cumulative incidence of disease progression

Patterns of co-mutation

• SF3B1 mutations define a distinct molecular subgroup that is mutually exclusive tomutations in TP53, and / or cytogenetic abnormalities or complex karyotype.

• Co-mutation with DNMT3A (20%) : - No effect on overall survival or event free survival - Increased involvement of multi lineage dysplasia

• Co-mutation with ASXL1 (7%), RUNX1 (6%) , EZH2 (5%) : - Increased transfusion dependency - Worse prognosis ( P= 0.004), disease progression (P=0.002)

Larger numbers are warranted to study the effects of co-mutation in MDS prognosis

SF3

B1

:MD

S

Malcovati et al Blood, 2015

SF3

B1

:MD

S

• Defines a distinct clinical and pathologic entity, one characterised by the presence of ringed sideroblasts;

• Molecular testing for SF3B1, more accurately identifies patients in this group, compared to %blast / % ringed sideroblast thresholds;

• SF3B1 is an independent prognostic predictor of clinical outcome associated with favorable prognosis;

• Attention should be given on the pattern of co-mutation: ASXL1 : RUNX1: EZH2 and potentially others..

SRSF

2:M

DS • Hotspot mutations, P95, Challenging molecular region, low NGS coverage

sometimes misses mutations.

• High risk MDS & CMML & sAML

• Associated with granulocytic disease, poor outcome and increased risk of leukemic transformation.

• Poor prognosis irrespective of estimates clonal status.

• Is frequently co-mutated with other adverse prognosis genes: ASXL1, STAG2, RUNX1, resulting in further deterioration in overall survival.

SRSF

2:M

DS • Hotspot mutations, P95, Challenging molecular region, low NGS coverage

sometimes misses mutations.

• High risk MDS & CMML & sAML

• Associated with granulocytic disease, poor outcome and increased risk of leukemic transformation.

• Poor prognosis irrespective of estimates clonal status.

• Is frequently co-mutated with other adverse prognosis genes: ASXL1, STAG2, RUNX1, resulting in further deterioration in overall survival.

SRSF

2:M

DS • Hotspot mutations, P95, Challenging molecular region, low NGS coverage

sometimes misses mutations.

• High risk MDS & CMML & sAML

• Associated with granulocytic disease, poor outcome and increased risk of leukemic transformation.

• Poor prognosis irrespective of estimates clonal status.

• Is frequently co-mutated with other adverse prognosis genes: ASXL1, STAG2, RUNX1, resulting in further deterioration in overall survival.

SRSF

2:M

DS • Hotspot mutations, P95, Challenging molecular region, low NGS coverage

sometimes misses mutations.

• High risk MDS & CMML & sAML

• Associated with granulocytic disease, poor outcome and increased risk of leukemic transformation.

• Poor prognosis irrespective of estimates clonal status.

• Is frequently co-mutated with other adverse prognosis genes: ASXL1, STAG2, RUNX1, resulting in further deterioration in overall survival.

SRSF

2:M

DS

U2

AF1

:MD

S • Hotspot mutations,

• ~ 7%, mostly high risk MDS & CMML & AML

• Associated with young age, and high-risk of leukemia transformation.

• Adverse prognosis:

Wu et al Am J Hematology 2014; Thol et al Blood 2012

Pro

gno

stic

eff

ect

of

splic

ing

fact

or

mu

tati

on

s

• Whilst SF3B1, associated with favourable prognosis;

• SRSF2; U2AF1, associated with inferior outcomes;

• ZRSR2 – unclear:

MD

S w

ith

sp

licin

g fa

cto

r m

uta

tio

ns

SF3B1

SRSF2 : U2AF1 ZRSR2 SF1 PRPF40B

Clinical presentation and clinical outcomes

Splic

ing

fact

or

mu

tati

on

s in

mye

loid

neo

pla

sms

Positive testing for splicing factor mutations is not unique to MDS

Splic

ing

fact

or

mu

tati

on

s in

mye

loid

neo

pla

sms

SRSF2 has shared genomic landscape between MDS and AML

Splic

ing

fact

or

mu

tati

on

s in

mye

loid

neo

pla

sms

Papaemmanuil under review, Lavallée Blood 2015

SF3B1 mutations occur in distinct genetic background in AML

Dia

gno

stic

cri

teri

a Testing positive for mutations in splicing factor genes..

Busque et al 2012; Jaiswal et al NEJM 2014; Genovese et al NEJM 2014, Xie et al Nat Medicine 2014; McKerrel et al Cell reports

Mu

tati

on

acq

uis

itio

n

Busque et al 2012; Jaiswal et al NEJM 2014; Genovese et al NEJM 2014, Xie et al Nat Medicine 2014; McKerrel et al Cell reports

Mu

tati

on

acq

uis

itio

n

Clin

ical

pre

sen

tati

on

• Presentation of splicing factor mutations in myeloid neoplasms

% of patients with mutation in panel

21

75

mye

loid

neo

pla

sms

Data presented from panel sequencing

Clin

ical

pre

sen

tati

on

• Presentation of splicing factor mutations in myeloid neoplasms

% of patients with mutation in panel

21

75

mye

loid

neo

pla

sms

Data presented from panel sequencing

Clin

ical

pre

sen

tati

on

Imp

ort

ant

con

sid

erat

ion

s

Formally modeled genomic structure to account for:

Gene mutations + genetic interactions + diagnostic variables + demographic variables.

and build personalised prediction models

26 centers: 2600 cases committed

Meta-analysis ~5000-6000 well annotated MDS cases

2015

Today

Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Invitation to Join Project

Begin receiving clinical data

Finish receiving clinical data

Send out sample queries and MTAs

Begin processing MTA agreements

Begin sending sample tubes

Begin receiving samples at MSKCC

QC and library Construction

Ack

no

wle

dge

me

nts

Peter Campbell Wellcome Trust Sanger Institute• Moritz Gerstung

AML- SG • Hartmut Dohner• Richard Schlenk • Lars Bullinger • Konny Dohner

Papaemmanuil Lab / MSKCC Leukemia Genomics • Gunes Gundem• Matahi Moarii• Franck Rappaport• Juan Medina• Komal Rathi • Noushin Farnhoud • Minal Patel• Kristina Knapp• Irene Phillip• Yesenia Werner• Marc Robert de Massy

MSKCC iGO• Agnes Viale• Kety Huberman